Rationale:
As of release 1.1.2[1], the configuration ini file supports
declaration of the `[main]` header as an alternative to global
properties by enumerating all sections and mapping each to the
respective parsing function. Global properties will still be parsed
correctly by fnott however generation adds unnecessary complexity to
the module. This commit removes the need for global properties
generation.
Changes:
- Fixed the FIXME at L118.
- Cleaned up unneeded let bindings.
- Changed the generation method to use the `pkgs.formats.ini` from
pkgs-lib instead of the raw `generators` library. This was done for
consistency and clarity as the `pkgs.formats.ini` is still required
for type declaration and uses `generators` internally.
- Removed `global-properties` testcase.
- Updated `example-settings` testcase.
[1] - https://codeberg.org/dnkl/fnott/releases/tag/1.1.2
* screen-locker: Make xautolock optional, reorganize options
xautolock isn't really needed to trigger xss-lock on the basis of time
since the built-in screensaver functionality of X serves as one of the
event sources for xss-lock. Keeping it around and defaulting to
"enabled" to avoid unexpected breakage.
Also shuffled around the options to submodules for xss-lock and
xautolock to get rid of prefixes in option names and to make
enableDetectSleep a bit clearer.
* screen-locker: update maintainership
* tests/screen-locker: Stub i3lock and xss-lock
* screen-locker: add package options for xss-lock and xautolock
kanshi configurations can have more than one exec statement in a
profile. This change allows services.kanshi.profiles.<name>.exec to be
a list of strings rather than a single string.
This option provides a more convenient way to overlay dummy packages.
It also adds a function `config.lib.test.mkStubPackage` that can,
e.g., be used for `package` options.
- Change generation behavior to always generate a configuration file
and pass it explicitly to fnott, it enforces the module to be
hermetic instead of offloading the configuration selection to
heuristics.
- Various style changes.
- Fix issue where fnott would abort due to an invalid config file when
both the configFile and settings options are unset.
- Remove the empty-settings test as a configuration file is now
already generated.
Suggested-by: Robert Helgesson <robert@rycee.net>
Fnott is a keyboard driven and lightweight Wayland notification daemon
for wlroots-based compositors.
There are four unit test to validate behavior for an empty
configuration, the default configuration, global properties and
systemd service file generation.
Fluidsynth's systemd unit currently has a hard dependency on the
pulseaudio systemd service. Since fluidsynth can use other sound
services (e.g., pipewire-pulse), this should be configurable. This
commit adds the relevant option.
PR #2238
* bspwm: various improvements
- fixes shell escaping issues and general style issues
- allow reloading the config on-the-fly by exposing bspwmrc to the user
* bspwm: add configuration test
When running a socket-activated emacs service, we don't want emacs to
remove the socket file after exiting, because then subsequent
invocations of `emacsclient` won't be able to use the socket to start
emacs.service again.
Emacs 27 added Type=notify support and updated the service definition to
remove the use of `emacsclient' to kill the service. Emacs 28 changes
the `StartupWMClass' in emacsclient.desktop to `Emacsd'. Update our
emacs.service and emacsclient.desktop definitions to match upstream
changes.
When killing emacs.service, the socket is removed, and subsequently
starting the service manually results in a service without a socket.
Prevent this by adding `RefuseManualStart=true' to the service's Unit
definition.
Drop Emacs 26 support as it is no longer shipped in nixpkgs. Update the
tests to verify the following configuration scenarios:
- Emacs version: 27, 28
- Socket activation: disabled, enabled
This change makes the services created via the lieer module aware of the notmuch config created by the home-mangager notmuch module (which is stored in a non-standard location).
Without this change all the lieer services created by the lieer module failed for me, as they were unable to find the notmuch config.
* i3, sway: extract border functionality to common function
Converted the i3 module to use default_border and
default_floating_border and extracted that functionality out to be
shared between the i3 and sway modules.
* i3: add sumnerevans as maintainer
Polybar's config format is a bit strange, and lists in particular are
annoying to handle. This enables using normal nix lists and nested
attrsets instead.
This change is not backwards-compatible, because the INI converter
converts lists of strings to space-separated values, and this does
something else. I expect that this is only relevant for the
`modules-left` etc bar setting, but that's enough to break things :(.
Not every option is exposed by redshift/gammastep parameters, for
example gamma options are only exposed in configuration file. So this
PR refactors this module to generate a configuration file and pass it
to the redshift/gammastep using -c parameter.
This is a breaking change since there is no support for some of the
older options like `extraOptions`, but unless you use `extraOptions`
it should work without changes.
Not every option is exposed by redshift/gammastep parameters, for
example gamma options are only exposed in configuration file. So this
PR refactors this module to generate a configuration file and pass it
to the redshift/gammastep using -c parameter.
This is a breaking change since there is no support for some of the
older options like `extraOptions`, but unless you use `extraOptions`
it should work without changes.
When setting `...sway.package = null`, the default bar configuration
would throw an error trying to use the bar from the null package.
Logic is added to use the bar from `pkgs.sway` instead of `cfg.package`
if it is null.
Fixes#1714
This adds a "test.asserts" module that currently just provides a
convenient way to assert on the content of warnings. By default all
tests will assert that no warnings are given.
Using the final package in the `onChange` block broke some use cases.
This restores the old behavior and instead solves the test
dependencies in a different way.
Fixes#1611
This reverts commit 7c3c64208e.
Emacs populates 'exec-path' at launch from the 'PATH' environment
variable. Likewise, the emacs derivation from nixpkgs populates
'load-path' from the 'NIX_PROFILES' variable. As neither of these are
available by default in the systemd user manager, revert to the
previous behavior of launching the Emacs daemon from a login shell.
Fixes#1354Fixes#1340
PR #1355
Add 'services.emacs.socketActivation.enable' for generating an
'emacs.socket' systemd unit.
Emacs since version 26 has supported socket activation, whereby an
external process manager such as systemd listens on a socket and passes
it to the Emacs daemon when the manager launches it. This improves
startup time of the user session and avoids launching the daemon when not
needed, for example when launching the user session via SSH.
This implementation hard-codes the socket path to the default for the
version of 'programs.emacs.finalPackage', because systemd does not
perform shell expansion in the socket unit's 'ListenStream' parameter
and it seems like an advanced use-case to change the socket path. Shell
expansion would be desirable as the socket path usually resides in
directories such as $XDG_RUNTIME_DIR or $TMPDIR.
Tests were added to verify behavior in the following cases:
- Emacs service with socket activation disabled
- Emacs 26 with socket activation enabled
- Emacs 27 with socket activation enabled
PR #1314
Before the XDG variables would be set from the user's environment, if
available. This would break some tests.
With this change the tests should be fully deterministic.
Fixes#1190
Add 'services.lieer', which generates systemd timer and service units
to synchronize a Gmail account with lieer. Per-account configuration
lives in 'accounts.email.accounts.<name>.lieer.sync'.
This replaces some derivation outputs by simple strings rather than
full Nix store paths. This removes the need to download the whole
derivation when all we need is a static string.
